In the first article in the series, “Any Fool Can Know. The Point is to Understand” we discussed how utilizing an Information Centric Approach can improve decision making capabilities. Now we will discuss how information presentation design fosters insight.
Many companies’ BI solution design still feature a very hierarchical, structured information presentation, which can disrupt people’s natural analytical methodology and hinder their path to insight. Rethinking the display methodology, to both prioritize the information, and to emphasize issue identification, validation and investigation significantly improves overall usability, and retained knowledge. This need to focus is further magnified, as users increasingly access and consume this information on mobile devices.
A mature, fully functional BI system is able to apply analytics to transform large amounts of data into information, providing insight and spawning action. As a Retail Practice Advisor, I have had the opportunity to work with a number of companies at different points along their BI maturity lifecycle, and have seen many struggle with efficiently transforming information into knowledge.
“Design is not just what it looks and feels like. Design is how it works” – Steve Jobs
To maximize the benefit of the BI analytics, information needs to presented in a way that serves not only to monitor and advise, but also enables intuitive investigation. The information presentation design needs to incorporate a detailed understanding of role-specific interaction and consumption. The design, over and above the normal 5 W’s (Who, Why, What, Where, When (and How)), “What next?” needs to be a crucial consideration.
Incorporating usage patterns, and their related data linkages and navigation, into the design also provides the capability to implement a proscriptive analysis and problem solving approach, by providing guided information access. While for some this runs counter to their ideas on how to maximize the capabilities of BI, many organizations desire a structured, methodical ‘best practices’ approach for certain roles.
Example – Let’s look at the concepts of visual design, information bundling and view navigation using an example of a Regional Operations Executive looking at a weekly sales report, and noticing that their St. Louis store’s sales were significantly down. How do you think the design would be different for Merchandising, Marketing, HR or Finance?
“If you can’t explain it simply, you don't understand it well enough” - Albert Einstein
Determining what information should be presented together, requires an understanding of the role-specific usage, along with the underlying data correlations. Providing the contextual data and causal factors that answer the initial logical questions, and also initiate a more comprehensive analysis, is a logical way to approach the bundling of information of any view.
In our example, some of the initial logical analysis paths would be: What about other Midwest stores? Are any specific departments causing this? How long has this been going on? What was the store traffic? Has competition changed? Were there weather issues? Prioritizing the ‘immediate’ questions that a user will have, provides an information bundling design focused on usability and productivity.
“Software is a great combination between artistry and engineering” – Bill Gates
Designing the visual presentation to both prioritize issue identification and provide overall insight, requires an understanding of the expected usage as well as people’s natural cognitive processes. This is a significant design challenge, because in addition to considering the most impactful way to present and visualize the data, there should be the flexibility to accommodate individual preferences. Some people are most comfortable with ‘data dense’ presentations, while others respond much better to a very visual presentation, with embedded cues, such as color / shape / movement.
Designing each presentation view to focus on the key metrics, and highlight exceptions and anomalies, requires a comprehensive understanding of the underlying data, since the exception could be based on predictive models, or peer groups, or trends and each of these suggests a different information presentation or visualization.
In our example, there are a number of visual models that would provide enhanced insight to the store’s sales. For a data centric person, sortable / filterable grids could provide peer rankings, and additional data fields could show other key fields such as customer count, average basket size, etc. A ‘visual person’ might prefer multivariate visualizations such as geo mapping (with weather data), scatter plots, heat maps etc., that identify key correlations, providing immediate understanding and identifying further analysis paths.
“Determine the thing that can and shall be done, and then we shall find the way” Abraham Lincoln
The goal of “turning insight into action”, requires a slightly different design approach. Understanding the actions that might be the desired end result of investigation into any metric, leads to the design of a guided analytics enabled solution. Several BI packages now contain the ability to develop this through a ‘machine learning’ approach, based on user behavior.
Thinking of what can be acted upon, and how, and what information is required, is a useful counterpoint to a top-down approach to information bundling and view navigation, and the combination of these two approaches provides a more robust solution.
In our example, the identification of causal factors (if any), would dictate the action paths that could be taken.
For example, if the initial analysis revealed that a new competitor entered the market, the logical actions could be to increase promotional activity in the store, or adjust the competitive pricing profile.
The most productive BI solutions incorporate usage and data relationships into their information presentation design, with a defined focus on issue recognition and user response. Leonardo Da Vinci’s commentary that “Simplicity is the ultimate sophistication”, is a tremendous guideline for designing end user presentations that contribute to optimal insight, usability and results.